► Click here to download the original poster (PDF)
Parag Dharapm, MD1, Sebastien Raimbault, PhD2
1 Dr. Dharap's Diagnostic Centre, Mumbai, India, 2 HORIBA Medical R&D, Montpellier, France
Though spread by different subspecies of vector insects, Malaria and Dengue are frequently found to be coexistent in economically challenged endemic area of developing countries. Common symptomatology poses challenges to health care givers. Encouraged by the performance of a previous malaria flagging algorithm applications developed through contemporary computer machine-learning techniques on blood cell counter, HORIBA decided to develop similar tools to screen for Dengue fever, as well as sub speciation of Malarial parasites, as a part of continual improvement endeavor. We evaluated in this study the performance of these flags on the HORIBA Yumizen H550 analyzer.
Uncontrolled population growth, ill planned urbanization, inadequate resources for public health & overall lack of awareness have possibly been the chief causes for the resurgence of one of the most common viral disease, Dengue Fever, in tropical region & developing nations. With heavy unpredictable monsoons it has become one of the leading infectious diseases causing public health menace.
Though spread by different subspecies of vector insects, Malaria and Dengue are frequently found to be coexistent in economically challenged endemic area of developing countries. Besides mosquitoes as vectors, both diseases share similar incubation periods in humans of 5 to 6 day, along with similar signs & symptomatology of acute febrile illness, with or without myalgia, hepatosplenomegaly & thrombocytopenia. Absence of other classical findings pose challenges to clinicians, who constantly desire accurate & quick differentiation between these major public health concerns for better therapeutic management.
Other than above mentioned common features, various studies have documented hematological parameter abnormalities like hemoglobin, hematocrit, MCHC, leukopenia, leukocyte population abnormalities, as well as biochemical parameter abnormalities including liver function tests.
Bhargava et al have suggested use of discriminant factors based on Volume, Conductivity and Scatter Properties of Leucocytes (VCS Technology) in Beckman Coulter analyzers for rapid diagnosis of Malaria & Dengue. Even after introduction of Dengue & Malaria PCR technology on a global platform its requirements of special instrumentation, time consumption & economic constraints have kept them out of reach of common individuals. The gold standard method of malarial parasite detection by peripheral blood smear examination, although most commonly utilized & economical, is known to require technical expertise and may need repetitive testing, thereby increasing turnaround time. Similarly for Dengue infection, screening tests detecting presence of disease specific antigen &/ or antibody testing, preferably by Elisa technique, requiring additional instrumentation, are being used as the mainstay for therapeutic interventions.
Encouraged by the performance of a previous malaria flagging algorithm applications developed through contemporary computer machine-learning techniques on blood cell counter ABX Pentra XLR (a five part differential counter with Reticulocyte counting technique) & Microsemi CRP (a three part differential blood cell counter with CRP), HORIBA decided to develop similar tools to screen for Dengue fever, as well as assess possibility to carry out sub speciation of Malarial parasites, as a part of continual improvement endeavor. We evaluated in this study the performance of these flags on the HORIBA Yumizen YH550 analyzer, a new platform for the low and middle range market.
A total number of 1436 patients, who visited Dr. Dharap's Diagnostic Centre, Mumbai, India for routine complete blood count examination, during July to October 2018 were studied. Residual blood specimens were serially analyzed on HORIBA Yumizen H550 analyzer. Local ethical standards were followed; however, informed consent was not required, as study was done by anonymizing retained blood samples without any extra charges & not communicating results of the study to the clinician, so as to affect the diagnostic and therapeutic management of the patient. Besides the primary patient selection criteria of fever, diagnostic microscopic blood smear examination for malarial parasite and confirmation by antigenic testing for Malaria and Dengue NSl antigen were performed for all cases.
For screening of Malaria cases, Rapid Malaria Antigen detection test kits manufactured by SD Biosensor Healthcare Pvt. Ltd., Gurugram, India using monoclonal anti-P. falciparum HRP-11 (0.75+/- 0.15 µg) with monoclonal anti-P. vivax pLDH(0.75+/- 0.15 ug) respectively to detect presence of P. falciparum & P. vivax related antigen. For screening & diagnosis of Dengue, RecombiLISA NSl Antigen test utilizing pairs of specific polyclonal & monoclonal anti-dengue antibodies of all four serotypes (DENl, 2, 3, 4) and analytical sensitivity of 0.3ng/ml for type 2 NSl antigen, manufactured by CTK Biotech, Inc., United States of America was used.
All the samples were identified & recruited for evaluation with screening/ diagnostic testing & microscopic review by an experienced pathologist. Samples were grouped as Normal, Malaria positive & Dengue positive cases. Malaria cases were further classified by high magnification microscopic review of 2,000 red cells by an experienced pathologist for speciation, dominant life cycle forms and parasite density. Persons for routine health check-up without history of any specific ailments were considered for the Normal or control group {N = 60). Samples were analyzed (N=1424, Vivax Positive N=184, Falciparum Positive N=21, Dengue fever N=213, other Pyrexia N=933) with prototype integrated instrument software comparing flagging results obtained with confirmatory testing.
Many diseases affect different blood cell parameters in direct or indirect ways resulting in detectable or undetectable changes in blood cell counts. Different studies have suggested combinations of different instrument measured parameters & indices for improving the efficiency in the process of malaria diagnosis. Computer assisted machine learning techniques considers a composite of all such measured parameters to identify and leverage those parameters of greater distinguishing power & significance.
Results of this study indicate that with the help of computer assisted machine learning techniques, it is possible to develop an effective, economical and helpful tool (suspect flags) to predict, screen and diagnose specific 'signatures' of the disease, as malaria and dengue. Insufficiency of Falciparum positive cases has been a limitation of this study. Larger multisite validation study may bring further insight & additional discerning power to this cost efficient technique with results comparable to the costlier technologies like light depolarisation, DNA/RNA fluorescence & light scatter signals, cell population changes used in other analysers.
Do you have any questions or requests? Use this form to contact our specialists.